# Conquistador

> Conquistador helps your coding agent do growth, marketing, sales, product and knowledge work. Give it an outcome and the facts; it picks from 38 methods and returns one reviewed draft. Private alpha.

Status: private alpha 0.0.14. Access by invitation. Public alpha starts at 0.1.0; no date yet.

Request access: [hi@hungv47.com](mailto:hi@hungv47.com?subject=Conquistador%20private%20alpha)

## What it is

Conquistador helps your coding agent do growth, marketing, sales, product and knowledge work. You give it an outcome and the relevant facts. It selects from 38 methods, names the specialists it uses, and returns one reviewed draft with review findings. Your coding agent supplies the model and tools.

## How it works

1. Install the CLI once. Private alpha: access by invitation. Invited accounts install a pinned private release with Node 24 and npm. There is no public registry package yet.
2. Run `conquistador` in the project. Guided setup picks the project and host, shows the plan and asks once. It adds no dependency or lockfile and starts no daemon, watcher or schedule.
3. Ask for an outcome in a fresh session, for example: "Use Conquistador to draft a launch plan from the product facts in this project. Mark missing facts. Keep it as a draft."

Setup writes:

```text
.conquistador/      
  SKILL.md          Start here
  README.md         Usage and lifecycle
  library/          All 38 methods and their resources
  agent/agent.json  Portable operator contract
  hosts/            BB specialist adapter
```

Commands:

- `conquistador`: Guided setup for this project
- `conquistador start`: Show the skill location and first task again
- `conquistador skills`: Browse the available capabilities
- `conquistador doctor`: Check local files, routing and hook registration
- `conquistador route --prompt '…'`: Preview a route before you run it
- `conquistador uninstall`: Remove the unchanged owned copies

## Methods (38)

### Growth and launch (13)

- Shape an ambiguous initiative (`shape-initiative`)
- Research market and positioning (`research-positioning`)
- Research content ideas (`research-content-ideas`)
- Diagnose a growth problem (`diagnose-growth`)
- Prioritize opportunities (`prioritize-opportunities`)
- Model a growth or revenue target (`model-growth-funnel`)
- Allocate a marketing budget (`allocate-marketing-budget`)
- Design pricing and packaging (`design-pricing-and-packaging`)
- Research or choose a channel (`research-channel`)
- Plan a campaign or launch (`plan-campaign`)
- Create paid media (`create-paid-campaign`)
- Create a run of show (`create-run-of-show`)
- Improve search and answer visibility (`optimize-search`)

### Marketing and sales (10)

- Establish a brand foundation (`create-brand`)
- Write a production brief (`brief-creative`)
- Write product and campaign copy (`write-copy`)
- Write social and community posts (`write-social`)
- Write direct outreach (`write-outreach`)
- Write long-form work (`write-longform`)
- Inspect a supplied video (`analyze-video`)
- Create short-form video (`create-shortform`)
- Improve a conversion surface (`improve-conversion`)
- Create or revise Vietnamese (`polish-vietnamese`)

### Product and engineering (6)

- Map a product journey (`map-user-flow`)
- Specify an interface (`brief-product-ui`)
- Architect a system (`architect-software-system`)
- Build an iOS experience (`build-ios-app`)
- Build a web experience (`build-web-app`)
- Write technical documentation (`write-technical-docs`)

### Knowledge and review (8)

- Learn from growth results (`measure-growth`)
- Evaluate paid results (`evaluate-paid-campaign`)
- Evaluate a sent outreach cycle (`evaluate-outreach`)
- Evaluate published short-form (`evaluate-shortform`)
- Audit a marketing package (`audit-marketing`)
- Review a finished artifact (`fresh-eyes-review`)
- Audit the facts behind a claim (`knowledge-review`)
- Resolve a consequential decision (`decision-panel`)

Plus one opt-in method: Share product feedback, opt-in (`submit-feedback`).

Specialist roles: Ads, Copy, Direct-response landing, SaaS landing, Data diagnosis, Campaign data, Creative assets.

## Coding agents

- Codex: `.agents/skills/conquistador`. Native project skill.
- Claude Code: `.claude/skills/conquistador`. Native project skill.
- Cursor: `.cursor/skills/conquistador`. Native project skill.
- GitHub Copilot: `.github/skills/conquistador`. Native project skill.
- BB: `.conquistador/`. Complete operator and specialist adapter.
- Hermes Agent: `.hermes/skills/conquistador`. Via conquistador --bot hermes, then a separate trust step.
- Any other agent: read `.conquistador/SKILL.md`.

## Principles

- You steer every output.
- Drafts, not auto-posts. Publication, spend, sends, external writes and saved memory keep their human decision. Accepting a draft does not authorize saved memory or feedback disclosure.
- Your model, your tools, your permissions. Your coding agent supplies the model, context, tools and permissions. Credentials stay with the host or its connection manager.
- Nothing runs in the background. Installation starts no daemon, watcher, hook or schedule. Activation defaults to manual: it works when you ask.
- Evidence kept separate. A passing local check does not prove a good result. Conquistador reports install checks, observed behavior and your acceptance as separate things.

## FAQ

### What is Conquistador?

Conquistador helps your coding agent do growth, marketing, sales, product and knowledge work. You give it an outcome and the relevant facts. It selects from 38 methods, names the specialists it uses, and returns one reviewed draft with review findings.

### Which coding agents does it work with?

Codex, Claude Code, Cursor and GitHub Copilot get a native project skill. BB uses the complete operator with its specialist adapter. Hermes Agent uses conquistador --bot hermes and a separate trust step. Any other agent can read .conquistador/SKILL.md directly. Open a fresh session after setup to confirm your host discovers it.

### What does it cost, and what is the license?

The source carries the MIT License; during the private alpha, source access is by invitation. Conquistador does not include a model: your coding agent supplies the model and tools, so model usage stays on the plan you already have with your host.

### What does “private alpha” mean?

Version 0.0.14 is a private 0.0.x release. The repository is private and access is by invitation. Invited accounts install a pinned, checksum-verifiable release from the private Git repository. No package is published to the npm registry.

### When is the public alpha?

The public alpha starts at version 0.1.0. There is no date yet. Its registry install command stays unannounced until the exact public package and a clean install lifecycle are verified.

### What does setup change in my project?

Guided setup writes a .conquistador folder and, for your host, one project skill folder. It adds no package manifest, dependency or lockfile to your project. conquistador uninstall removes the unchanged owned copies.

### Where do my data and credentials go?

They stay with your host. Work from supplied facts needs no connected account. When a task needs live access, Conquistador helps you connect only that service, and the credential stays with the host or connection manager. There is no automatic transcript collection or background feedback upload.

## About

Conquistador is a [FORSVN](https://forsvn.com) studio product from Buon Ma Thuot, Vietnam. Contact: hi@hungv47.com. Canonical page: https://conquistador.forsvn.com/